Inventory Forecasting for Bidded Ad Exchange

1. A non-transitory computer-readable medium having at least one computer-executable program stored thereon that, when executed in at least one computing device, causes the at least one computing device to perform operations comprising: obtaining data for a plurality of historical bid requests for a plurality of historical time periods from a data store, each of the plurality of historical bid requests indicating a past opportunity to submit a bid to present content in conjunction with a particular space of a user interface rendered on a network page, the past opportunity to submit the bid comprising a bid request received from an exchange device that operates an electronic bidding exchange, the exchange device transmitting the bid request in response to obtaining from a network content device a content call that indicates an availability of the particular space in the user interface of the network page; determining a plurality of historical bid request quantities for each of the plurality of historical time periods, each of the plurality of historical bid request quantities corresponding to at least one of a plurality of content space characteristics; determining a plurality of expected historical bid request quantities for each of the plurality of historical time periods, each of the plurality of expected historical bid request quantities corresponding to the at least one of the plurality of content space characteristics; determining a plurality of weighting values for the plurality of historical time periods based at least in part on a comparison between the plurality of historical bid request quantities and the plurality of expected historical bid request quantities; selecting a quantity of the plurality of historical bid requests for each of the plurality of historical time periods, each quantity being selected based at least in part on a respective one of the plurality of weighting values; associating a plurality of booked impressions for a plurality of advertising campaigns with the plurality of historical bid requests that are selected, wherein the plurality of booked impressions are associated based at least in part on a plurality of criteria associated with the plurality of advertising campaigns and comprise at least one obligation for the at least one computing device to cause at least one instance of the content to be served in conjunction with the user interface rendered on the network page, the plurality of criteria comprising at least one of a geographical location of a respective client device or a size of the particular space in the user interface of the network page; and determining an expected bid requests inventory based at least in part on the plurality of booked impressions that are associated with the plurality of historical bid requests and a determination of a quantity of the historical bid requests remaining after a subset of the historical bid requests are associated with the plurality of booked impressions, the expected bid requests inventory representing an expected quantity of bid requests that will be received and that would satisfy the plurality of criteria for presenting the content on the user interface of the network page.

2. The non-transitory computer-readable medium of claim 1 , wherein the operations further comprise determining an average expected bid requests inventory based at least in part on a plurality of iterations of the plurality of booked impressions being associated with the plurality of historical bid requests that are selected.

3. The non-transitory computer-readable medium of claim 2 , wherein the operations further comprise: determining a plurality of bid success rates for a plurality of historical bid amounts; and determining a plurality of expected impressions inventories for a proposed advertising campaign, the plurality of expected impressions inventories being based at least in part on the average expected bid requests inventory and each of the plurality of bid success rates.

4. The non-transitory computer-readable medium of claim 1 , wherein the plurality of criteria comprise a plurality of target advertising segments and a location of the particular space with respect to the network page.

5. The non-transitory computer-readable medium of claim 1 , wherein the quantity of the plurality of historical bid requests is a one-dimensional array, wherein an element of the one-dimensional array corresponds to the quantity of the plurality of historical bid requests that were received.

6. The non-transitory computer-readable medium of claim 1 , wherein the plurality of criteria comprise a frequency cap for presenting the content to a respective client device.

7. The non-transitory computer-readable medium of claim 1 , wherein the content comprises a Hypertext Markup Language (HTML) snippet that is used to render the content on the network page.

8. A system, comprising: at least one computing device; and a forecasting engine executable by the at least one computing device, the forecasting engine configured to cause the at least one computing device to: obtain data for a plurality of historical bid requests for a plurality of historical time periods, each of the plurality of historical bid requests indicating a past opportunity to submit a bid to present content in conjunction with a user interface rendered on a network page; generate a forecasting model to determine an expected bid requests inventory, the forecasting model using a selected quantity of the plurality of historical bid requests for each of the plurality of historical time periods, wherein the selected quantity for each of the plurality of historical time periods is based at least in part on a respective weighting value of a plurality of weighting values for the plurality of historical time periods; determine a plurality of historical bid requests quantities for each of the plurality of historical time periods, each of the plurality of historical bid requests quantities corresponding to at least one of a plurality of characteristics; determine a plurality of expected historical bid requests quantities for each of the plurality of historical time periods, each of the plurality of expected historical bid requests quantities corresponding to the at least one of the plurality of characteristics; determine the respective weighting value for each of the plurality of historical time periods based at least in part on a comparison between the plurality of historical bid requests quantities and the expected historical bid requests quantities; and determine an expected bid requests inventory based at least in part on the forecasting model and a determination of a quantity of the historical bid requests remaining after a subset of the historical bid requests are associated with a plurality of booked impressions, the expected bid requests inventory representing an expected quantity of bid requests that will be received and that would satisfy a plurality of criteria for presenting the content on the user interface of the network page.

9. The system of claim 8 , wherein the forecasting engine further causes the at least one computing device to: determine a plurality of bid success rates for a plurality of historical bid amounts; and determine an expected impressions inventory for a content campaign, the expected impressions inventory being based at least in part on an average of the expected bid requests inventory and at least one of the plurality of bid success rates.

10. The system of claim 8 , wherein the forecasting engine further causes the at least one computing device to: select the selected quantity of the plurality of historical bid requests for each of the plurality of historical time periods based at least in part on the respective weighting value corresponding to a respective one of the plurality of historical time periods; and wherein the plurality of historical bid requests are associated with the plurality of booked impressions based at least in part on at least one of the plurality of criteria.

11. The system of claim 10 , wherein the criteria comprise a plurality of target advertising segments.

12. The system of claim 10 , wherein the criteria comprise a limit on a serving quantity within a predefined time period.

13. The system of claim 10 , wherein the criteria comprise a predefined time period when the content is to be presented.

14. The system of claim 10 , wherein the criteria comprise a predefined geographical location where the content is to be presented.

15. The system of claim 8 , wherein at least one of the plurality of historical time periods is selected for being a day of a week that is the same as the day of the week that the forecasting engine is executed.

16. The system of claim 15 , wherein the at least one of the plurality of historical time periods is selected for being the day of the week that is closest to a year prior to the day that the forecasting engine is executed.

17. The system of claim 15 , wherein the at least one of the plurality of historical time periods is selected for being associated with a holiday that is associated with a time period when the forecasting engine is executed.

18. A method, comprising: obtaining, in at least one computing device, data for a plurality of historical bid requests for a plurality of historical time periods, each of the plurality of historical bid requests indicating a past opportunity to submit a bid to present content in conjunction with a user interface; determining, in the at least one computing device, a plurality of historical bid requests quantities for each of the plurality of historical time periods, each of the plurality of historical bid requests quantities corresponding to at least one of a plurality of characteristics; determining, in the at least one computing device, a plurality of expected historical bid requests quantities for each of the plurality of historical time periods, each of the plurality of expected historical bid requests quantities corresponding to the at least one of the plurality of characteristics; generating, in the at least one computing device, a forecasting model based at least in part on a selected quantity of the plurality of historical bid requests for each of the plurality of historical time periods, each selected quantity being based at least in part on a respective weighting value of a plurality of weighting values that correspond to the plurality of historical time periods; determining, in the at least one computing device, an expected bid requests inventory based at least in part on a determination of a quantity of the historical bid requests remaining after a subset of the historical bid requests are associated with a plurality of booked impressions, the expected bid requests inventory representing an expected quantity of bid requests that will be received and that would satisfy a plurality of criteria for presenting the content on the user interface of a network page; and determining, in the at least one computing device, the respective weighting value for each of the plurality of historical time periods based at least in part on a comparison between the plurality of historical bid requests quantities and the plurality of expected historical bid requests quantities.

19. The method of claim 18 , further comprising determining, in the at least one computing device, the expected bid requests inventory using the forecasting model.

20. The method of claim 19 , further comprising: determining, in the at least one computing device, a plurality of bid success rates for a plurality of historical bid amounts; and determining, in the at least one computing device, an expected impressions inventory for a content campaign, the expected impressions inventory being based at least in part on the expected bid requests inventory and at least one of the plurality of bid success rates.

21. The method of claim 20 , further comprising: determining, in the at least one computing device, the expected bid requests inventory using the forecasting model for a plurality of iterations; and determining, in the at least one computing device, an average expected bid requests inventory, wherein the expected impressions inventory is further based at least in part on the average expected bid requests inventory.

22. The method of claim 18 , wherein generating the forecasting model further comprises: selecting, in the at least one computing device, the quantity of the plurality of historical bid requests for each of the plurality of historical time periods based at least in part on the respective weighting value that corresponds to a respective one of the plurality of historical time periods; and wherein the plurality of historical bid requests are associated with the plurality of booked impressions based at least in part on at least one of the plurality of criteria.

23. The method of claim 22 , wherein the plurality of criteria comprise a plurality of target advertising segments.

24. The method of claim 22 , wherein the plurality of criteria comprise a predefined size of a space where the content is to be presented.

25. The method of claim 22 , wherein the plurality of criteria comprise a predefined time period when the content is to be presented or a predefined geographical location where the content is to be presented.

26. The method of claim 22 , wherein the plurality of criteria comprise a frequency cap for presenting the content to a respective client device.

27. The method of claim 18 , wherein at least one of the plurality of historical time periods is selected for being a day of a week that is the same as a day of the week that the method is performed.

28. The method of claim 27 , wherein the at least one of the plurality of historical time periods is selected for being the day of the week that is closest to a year prior to the day of the week that the method is performed.